The email sat in the inbox, mocking her. "Celebrated Author, Now Accused of Plagiarism." The headline screamed at her from the screen. Her stomach churned. It wasn’t a surprise, not really. The way the plot points in Elara Thorne’s new novel mirrored her own unpublished manuscript had been gnawing at her for weeks. Now, it was confirmed. She slammed the laptop shut, the metallic clang echoing in the otherwise silent apartment.

She paced. Her nails, usually meticulously painted, were now bitten to the quick. She grabbed her phone, scrolling through the contact list, but stopped before pressing dial. Her friends wouldn't understand. They’d say “move on,” and "it's not worth it." But her manuscript *was* worth it. It was her heart, her soul, poured onto the page. Thorne had stolen it, and she felt a heat rising in her chest.

She went to the kitchen, yanking open the fridge, searching for something to destroy. Found a jar of pickled onions. She hurled it across the room. The glass shattered, the pungent smell of vinegar filling the air. She stood, breathing raggedly, staring at the mess. Then, a slow smile crept across her face. This was just the beginning.

Emotion: hostile

Cluster: Anger / Hostility
PC1 (Valence): -1.67 Negative
PC2 (Disposition): 1.56

Role in Research

This story is one of 1,000 stories generated for the emotion hostile. During extraction, it was fed through Gemma4-31B and its hidden state activations were captured at 11 layers.

The mean activation across all 1,000 hostile stories, after denoising with neutral dialogue baselines, produces the hostile emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the hostile vector is projected through the unembedding matrix.

Promoted:
C0.592
S0.387
0.324
est0.301
aggravated0.295
Suppressed:
a-0.457
de-0.375
H-0.289
🤩-0.280
la-0.273